6129 from C-H bonds of alcohols to form radical pairs (e.g., PhCH2. C H 2 0 H ) which would recombine with high efficiency to give the C-H "insertion" products in the matrix environment, whereas in the liquid phase the members of the pairs would diffuse apart to yield 2 and 3. The facts that the yield of bibenzyl drops markedly as the reaction phase is changed from liquid to solid and 5 increases significantly in colder and presumably harder -196 "C matrixi0are apparently related to the relative diffusibility of radicals. The reason for the marked reduction in the ether yields a t lower temperature is uncertain. A rapid decay of the initially formed singlet carbene to the triplet in rigid matrix seems unlikely since styrene, an intramolecular product of singlet PhCMe," was formed in high yield even in -196 "C matrix and, further, rapid and reversible singlet-triplet equilibrium has been to be involved in phenylcarbene reactions. There is a possibility, however, that the 0 - H insertion reaction could be more significantly assisted by solvation, which might become less important in solid phase, compared with the other radical processes, since ionic species,la Le., benzyl cation and/or ylide, have been proposed to be involved in the reaction. Thus, matrix might affect the kinetics of the competitive intermolecular singlet and triplet carbene processes occurring within it.5 The low temperature photolysis3 of diphenyldiazomethane in tert-butyl alcohol matrix gave a completely different picture than irradiation in other alcoholic matrices. There was no sign of the C-H insertion product being formed; instead, there was formed azine (8), ethylene (9), and ethane (10)
+-
hv
Ph2CN2 6
+
Ph2CHO-t-Bu Ph2C=NN=CPh2 7 8 Ph2C=CPh2 Ph2CHCHPh2 9 10
+
T , "C
7
21 -72 -196
91.5 3 .O 0.8
+
Relative yields, % 8 9 2.5 56.6 26.6
0 26.9 12.6
10
0 13.5 0
all of which were only minor products (
